Yesterday we released Cancel and Replace transactions in SDRView Professional and this article will provide some detail on this.
The CFTC Part 43 Rule for Real-time Public Reporting of Swap Transaction data, allows for the Cancellation and Correction of previously disseminated transactions. It is inevitable that any system that publishes real-time messages, needs to have a mechanism to cancel and correct erroneous messages.
What we now show
The screenshot below shows that we now show the Cancel and Replace transactions as they are published.
So from above we can see that:
- A trade with Identifier 8197744 was canceled by a subsequent transaction with Identifier 8199100.
- And then replaced with Identifier 8199101 which changed the rate from 2.9 to 2.905, so half a basis point.
- Also a trade with Id 8197742 went through a similar half a basis point change.
- We then apply these changes in our aggregate results panel.
How important are these Corrections?
Well it depends on what you are looking at and whether the detail matters to you.
To some it may seem like noise and in-consequential.
To others it may be very important.
As an example lets look at USD Interest Rate Swaps, vanilla spot starting, reported on 9 December 2013 in SDRView Professional.
By using the selection filters we can see the following:
- There were 290 On SEF Cleared Swaps, with $26b notional and only 1 correct and replace with a 10Y rate being corrected from 2.908 to 2.902.
- However if we look at all USD IRS, whether On or Off or Cleared or Uncleared, we see that there were 541 trades with $49b notional and 61 Cancellation transactions and 49 Replacement transactions.
- As there are no On SEF and Uncleared, we can say that the quality of Off Facility reporting is much worse than that of On SEF.
- Not surprising given that the On SEF trades are electronically captured at point of execution and dissemination by the SEF platforms directly to the DTCC SDR.
- While the Off Facility trades are manually entered at some point after execution and disseminated by many different systems at reporting swap dealers to the DTCC SDR.
- Digging deeper into the 61 Cancellations we can observe that the larger number of cancellations indicate that there were trades that were just cancelled either because they were duplicates of trades already reported or that they did not need to be reported as exempt from mandatory reporting or that they have been cancelled but the replacement has not yet been received.
- Digging deeper into the 49 Replacements we can see that the majority of these have Identifiers that are very close to the original trade being replaced, which means that the replacement was received within 1 or 2 minutes of the original trade. Also many of the others are so minor as too be almost inconsequential e.g. a change in execution time by 15 secs, or the rate by 1/2 bps or a change to another field we only show on the trade ticket view e.g. Indication of Collateralisation from UC to PC; so Un-collateralised to Partially-collateralised.
- So while on the face of it the cancel and replace count might seem high, the reality is much better than that.
Any real-time dissemination message protocol needs cancel and replace messages.
DTCC SDR has always had cancel and correct messages.
We did not apply these in the past as our analysis showed that doing so would not make a significant difference to the information content.
Now we have many more active users, we think the level of detail is more important.
And it is our job to “dive deep” and “sweat the detail”, so our users can glean even more information.
So today if you use SDRView Professional you will see Cancel and Replace transactions (assuming you look hard enough).
Any questions, please see the ? Help in the application or email us at email@example.com